Redeploy node with updated release and keep data in place

Registered by Oleg S. Gelbukh on 2015-09-30

For every node role, there could be a data that needs to be preserved during the in-place upgrade procedure.

We need to redeploy a node in cluster after it is updated to the new version
* mark specific partitions, depending on role and volumes configuration
* implement driver in fuel-agent to preserve raw data on marked partitions
* modify library modules and tasks to properly assume existing data set

Blueprint information

Status:
Not started
Approver:
Fuel Core Team
Priority:
Undefined
Drafter:
Oleg S. Gelbukh
Direction:
Needs approval
Assignee:
None
Definition:
Review
Series goal:
Proposed for newton
Implementation:
Deferred
Milestone target:
milestone icon 10.0

Whiteboard

Gerrit topic: https://review.openstack.org/#q,topic:bp/upgrade-redeploy-node,n,z

Addressed by: https://review.openstack.org/238931
    Spec for upgrade node with redeploy

Gerrit topic: https://review.openstack.org/#q,topic:bp/partition-preservation,n,z

Addressed by: https://review.openstack.org/203639
    Support Ceph OSD devices with existing data set

(?)

Work Items

Dependency tree

* Blueprints in grey have been implemented.

This blueprint contains Public information 
Everyone can see this information.

Subscribers

No subscribers.